From: Ted K. <te...@op...> - 2005-09-29 17:16:56
|
On Wed, 2005-09-28 at 15:57 -0400, Tarus Balog wrote: > On Sep 27, 2005, at 6:30 AM, Chris Phillips wrote: > > > I have 900 Cisco 837's and 20 Cisco 1760's. I want to poll all > > available > > data from the 1760's but only minimal throughput and uptime on the > > 837's. Looking at the config files for snmp there doesn't appear to be > > any way to accomplish this... attributes available seem to be solely > > defined by what oid's exist and such, and from that angle all devices > > appear the same. so I either leave it all as is, and get a nasty IO > > overhead and no disk space, or i lose information of critical network > > nodes. > > I would borrow the threshold category in assets, set up two > collection schemas, one for "all" and one for "little" and then use > that category to put each system into the proper bucket. > > -T Can you elaborate on this? It would be awesome if you mean one can map a datacollection group to a collectd package :-) Regards, Ted |